Correlation of job-shop scheduling problem features with scheduling efficiency
作者：
Highlights：
• A set of 380 features are developed for a Job-Shop Scheduling Problem.
• Supervised machine learning and statistical methods used for feature evaluation.
• Features are evaluated for their correlation with optimal makespan.
• Features are used for classification of instances based on optimal makespan.
• Potential application to a real-world manufacturing example is demonstrated.
摘要
•A set of 380 features are developed for a Job-Shop Scheduling Problem.•Supervised machine learning and statistical methods used for feature evaluation.•Features are evaluated for their correlation with optimal makespan.•Features are used for classification of instances based on optimal makespan.•Potential application to a real-world manufacturing example is demonstrated.
论文关键词：Job-shop scheduling,Scheduling efficiency,Makespan prediction,Machine learning,Support vector machines
论文评审过程：Received 26 August 2015, Revised 8 June 2016, Accepted 8 June 2016, Available online 9 June 2016, Version of Record 20 June 2016.
论文官网地址：https://doi.org/10.1016/j.eswa.2016.06.014
